No abstract

The structure of the novel compound La2MoO5 has been solved from powder X-ray and neutron diffraction data and belongs to the tetragonal space group P4/m (no. 83) with a = 12.6847(3) Å and c = 6.0568(2) Å and with Z = 8. It consists of equal proportions of bioctahedral (Mo2O10) and square prismatic (Mo2O8) dimers, both of which contain direct Mo-Mo bonds and are arranged in 1D chains. The Mo-Mo bond length in the Mo2O10 dimers is 2.684(8) Å, while there are two types of Mo2O8 dimers with Mo-Mo bonds lengths of 2.22(2) and 2.28(2) Å. Although the average Mo oxidation state in La2MoO5 is 4+, the very different Mo-Mo distances reflect the fact that the Mo2O10 dimers contain only Mo(5+) (d(1)), while the prismatic Mo2O8 dimers only contain Mo(3+) (d(3)), a result directly confirmed by density function theory calculations. This is due to the complete disproportionation of Mo(4+), a phenomenon which has not previously been observed in solid-state compounds. La2MoO5 is diamagnetic, behavior which is not expected for a nonmetallic transition-metal oxide whose cation sites have an odd number of d-electrons. The resistivity displays the Arrhenius-type activated behavior expected for a semiconductor with a band gap of 0.5 eV, exhibiting an unusually small transport gap relative to other diamagnetic oxides. Diffuse reflectance studies indicate that La2MoO5 is a rare example of a stable oxide semiconductor with strong infrared absorbance. It is shown that the d-orbital splitting associated with the Mo2O8 and Mo2O10 dimeric units can be rationalized using simple molecular orbital bonding concepts.

Among oxide compounds with direct metal-metal bonding, the YMoO (ABO) structural family of compounds has a particularly intriguing low-dimensional structure due to the presence of bioctahedral BO dimers arranged in one-dimensional edge-sharing chains along the direction of the metal-metal bonds. Furthermore, these compounds can have a local magnetic moment due to the noninteger oxidation state (+4.5) of the transition metal, in contrast to the conspicuous lack of a local moment that is commonly observed when oxide compounds with direct metal-metal bonding have integer oxidation states resulting from the lifting of orbital degeneracy typically induced by the metal-metal bonding. Although a monoclinic C2/m structure has been previously proposed for LnMoO (Ln = La-Lu and Y) members of this family based on prior single crystal diffraction data, it is found that this structural model misses many important structural features. On the basis of synchrotron powder diffraction data, it is shown that the C2/m monoclinic unit cell represents a superstructure relative to a previously unrecognized orthorhombic Immm subcell and that the superstructure derives from the ordering of interchangeable MoO and LaO building blocks. The superstructure for this reason is typically highly faulted, as evidenced by the increased breadth of superstructure diffraction peaks associated with a coherence length of 1-2 nm in the c* direction. Finally, it is shown that oxygen vacancies can occur when Ln = La, producing an oxygen deficient stoichiometry of LaMoO and an approximately 10-fold reduction in the number of unpaired electrons due to the reduction of the average Mo valence from +4.5 to +4.05, a result confirmed by magnetic susceptibility measurements. This represents the first observation of oxygen vacancies in this family of compounds and provides an important means of continuously tuning the magnetic interactions within the one-dimensional octahedral chains of this system.

The absolute absorption coefficient, α(E), is a critical design parameter for devices using semiconductors for light harvesting associated with renewable energy production, both for classic technologies like photovoltaics and for emerging technologies such as direct solar fuel production. While α(E) is well-known for many classic simple semiconductors used in photovoltaic applications, the absolute values of α(E) are typically unknown for the complex semiconductors being explored for solar fuel production due to the absence of single crystals or crystalline epitaxial films that are needed for conventional methods of determining α(E). In this work, a simple self-referenced method for estimating both the refractive indices, n(E), and absolute absorption coefficients, α(E), for loose powder samples using diffuse reflectance data is demonstrated. In this method, the sample refractive index can be deduced by refining n to maximize the agreement between the relative absorption spectrum calculated from bidirectional reflectance data (calculated through a Hapke transform which depends on n) and integrating sphere diffuse reflectance data (calculated through a Kubleka-Munk transform which does not depend on n). This new method can be quickly used to screen the suitability of emerging semiconductor systems for light-harvesting applications. The effectiveness of this approach is tested using the simple classic semiconductors Ge and Fe 2 O 3 as well as the complex semiconductors La 2 MoO 5 and La 4 Mo 2 O 11 . The method is shown to work well for powders with a narrow size distribution (exemplified by Fe 2 O 3 ) and to be ineffective for semiconductors with a broad size distribution (exemplified by Ge). As such, it provides a means for rapidly estimating the absolute optical properties of complex solids which are only available as loose powders.